Orosius may

6, 8, Eutropius, 5, 3, 1, and Orosius, 5, 18, I. have copied Eutropius, of whom, however, Florus was not the source, for he does not have the praenomina there given. It is likely that Florus and Eutropius represent the original order of the Epitome, some descendant or Ms of which changed it before the copies were made by Obsequens and Cassiodorus. That a different order stood in the Fasti was the reason for the change.

694 a.u.c. Obsequens, 62, and Cassiodorus agree against the other Fasti, Dio, 37, 49, 1, and C.I.L. I, 727–728.

572 a.u.c. Cassiodorus, Obsequens, 5, and Oxyr. 67 change the order of names from Livy, 39, 56, to agree with the Fasti, etc. I have shown above that the agreement of Hydatius with Livy for this year was due to the Epitome as intermediate source. We have also seen how rigidly in all cases the Epitome preserved the Livian order of names. The natural explanation, therefore, is that the immediate ancestor of Cassiodorus, Obsequens, and Oxyr. changed the order so as to agree with the Fasti.

I add certain Ms variations in spelling, which point to a similar near relationship.

For 578 a.u.c. Cassiodorus has Petillius as Fasti Capitolini and Obsequens, 9, but Livy, 41, 14, Petilius, with one l This has little weight owing to poor Ms authority of Livy, 41 to 45.

For 591 a.u.c. Obsequens, 14, has T. Graccho; Cassiodorus, T. Sempronius; Fasti Capitolini, Ti. Sempronius Gracchus II. Again, for 577 a.u.c. Cassiodorus has T. Sempronius, but the Fasti, Ti. Also Per. Liv. 41, and Orosius, 4, 20, 32, have Tiberius, so the Epitome must have had Tiberius, as also Livy, though T. Sempronius appears in the sole Ms of Livy, 41, 8, I.

For 650 a.u.c. Obsequens, 43, has Caio Marcio Caio Flacco; Cassiodorus, C. Marcius, C. Fl. Fimbrius. C.I.L. X, 3780, gives [C. Flavio C. f. C. Mario C. f. cos. Furthermore, the Fasti and all the descendants of the Epitome, including Cassiodorus and Obsequens in other passages, have the name of the famous C. Marius right. This is a Ms error of the parent of Obse

quens and Cassiodorus. In the second name it arose from the abbreviation C. Fl. Fimbr., which is still distinguishable in Cassiodorus.

For 652 a.u.c. the error of Obsequens, 44, Q. Luctatio (for Lutatio), appears in Cuspinian's edition of Cassiodorus, though the Regensburg Ms (11th cent.) has L. Lutatius. Mommsen (Leip. Akad. VIII, p. 571 ff.) has shown that Cuspinianus had the parent Ms of Cassiodorus before him.

For the years 723 to 731 a.u.c. Cassiodorus numbers the consulships of Augustus II to X instead of III to XI as the other authorities, including Orosius, 6, 19, 14; 6, 20, 1; 6, 21, I. The change had thus not taken place in the Epitome. It agrees better with the marks of revision which we have referred to the parent of Cassiodorus than with Cassiodorus himself.

From the evidence given by me (U. of M. Studies, I, p. 186) to show Ms variations in the original Epitome of Livy, the common mistake of Obsequens, 49, and Cassiodorus, 658, is best referred to the parent of these two. The correct form for the original Epitome is shown by Per. Liv. 70, Hieronymus, 1922, Festus, 13, etc.

In the following two cases, as in possibly one above, the variations may even go back to Livy himself, though it seems easier to refer them to a later writer.

For 624 a.u.c. Cassiodorus and Obsequens, 28, have App. Claudius, but Cicero, de Leg. 3, 19, 42, C. Claudius. For 659 a.u.c. Cassiodorus and Obsequens, 50, have P. Crassus, but Fasti Capitolini, L. Licinius Crassus.1

Summing up our evidence, we find that we have fifteen errors or changes, which are best explained by supposing a close relationship of source between Cassiodorus and Obsequens and nothing against this assumption. Considering the different character of the two and the brevity of the portion

1 For the year 576 a.u.c., Ay, p. 53, refers to the Epitome the mistake of Cassiodorus and Obsequens, 8, Cn. Manlius for A. Manlius of Livy, 40, 59. But the Livy passage has to be emended by reference to Livy, 41, 10; 43, 2; 45, 9, and also the form Cn. Manlius appears in Florus, 1, 26, 2, hence in the original Epitome.

of Obsequens which comes in question, this is a remarkable agreement.

It remains to compare this evidence of the consuls' names with the previous conclusions of a close agreement between Obsequens, Oxyr., and Per. I a in language and the chronological agreement of Hydatius and Cassiodorus. It is evident that most of our evidence belongs under the head of Ms variation, though cases of interpolation from other Mss or works are not infrequent. Many of the changes have been due to errors in writing, but we are not to think that this fact and the overwhelming majority of agreements in all the descendants preclude the possibility of an intermediate source for any of the later writers. We have seen that the Epitomator Livii followed Livy's consular list with even greater faithfulness. So while I have discussed these questions as Ms variations, and am convinced that we have the right to judge the source relationships of these few later writers as we would Mss and Ms families, this position is entirely compatible with the further conclusion that some of these families represent expansions, abridgments, or even excerpts from the original Epitome.

How many of these there have been we may never know; for the present I feel sure of only two:

(a) An abridgment used by Obsequens, the Oxyrhynchus Epitome, and the Periocha 1 a of Livy.

(b) The Livian consular list, with addition of some historical statements, excerpted from this abridged Epitome. It was used regularly by Cassiodorus and compared by the unknown author of the Hydatian Fasti, though his main source was the Fasti Consulares, which we can trace back through the source of the Fasti Capitolini (30 B.C.) to the Annales Maximi as final source.

We have a very few instances indicating that the Epitomator Livii was acquainted with these Fasti Consulares, which were in common use and were parallel to the consular names prefixed to each year in the Epitome of Livy. Owing to this common character and the ease of comparison, interpolations were liable to occur from one to the other.

Thus a few

changes and historical additions had come into the Fasti Consulares before 354 A.D., when the Chronograph made his copy. Later a second small series of interpolations came in which appear in Hydatius, though not drawn from the Livian Fasti, which were later excerpted from the abridged Epitome, and This show the influence of the chronology of Eutropius. work the Pseudo-Hydatius compared, when copying the Fasti Consulares. The following diagram shows the relationship of the various authors discussed.

The abridged Epitome of Livy must have been in existence before 300 A.D. It doubtless preserved the character and form of the original Epitome as described above. In size, if we may judge from the prodigies of Obsequens, it was much more voluminous than the Periochae of Livy.

The Fasti Liviani were excerpted after 375 and before 470 A.D. The consuls' names were taken practically in the form of the Epitome, though doubtless often shortened. If the consul had served before, the number of the consulship was noted. Though Cassiodorus more often omits these numbers, yet he preserves seven not found in Livy. The other descendants of the Epitome also preserve a few of the

numbers, Per. Liv. 101 even one not found in Livy. The filling out of the consular list in this manner was, therefore, the work of the Epitomator Livii, though some additional changes can be traced to the direct source of Cassiodorus, i.e. the excerptor of the Livian Fasti.

The names of the consuls were put in the ablative at the beginning of each year by the Epitomator, the lines projecting from the body of the text as in Oxyr. so that the consulships could be easily compared or counted. This form was retained in the abridged Epitome and so passed into Oxyr.

That the names were given in the ablative in the abridged Epitome is proved by the mistakes of Cassiodorus and the agreement in form of Oxyr. and Obsequens. This form is also just as sure for the original Epitome. Orosius gives the consuls' names in the ablative, introducing his statements for the year, more than fifty times, Eutropius nearly as many, and yet in only twenty-four cases do they give the consuls for the same years, in less than half of which Orosius copied Eutropius. All the rest of this long list came, as we have seen above, directly from the Epitome of Livy. The Per. Liv. also preserved this ablative form in six cases, though the work had ceased to be annalistic. Florus has still less reason for retaining this form, yet, to the injury of his style, he often gives the name of one consul in the ablative absolute and allows you to supply populus Romanus as subject of the sentence. A much worse case of awkward retention of the ablative absolute is found in Eutropius, 2, 22, 1: M. Aemilio Paulo Ser. Fulvio Nobiliore consulibus ambo Romani consules ad Africam profecti sunt. Such cases prove the presence of the ablative absolute in the original even more certainly than the frequent occurrence of the ablative in Eutropius and Orosius does.

We see thus that even the annalistic form and the designation of the years by the consuls' names appeared in the Epitome of Livy, so there was little or nothing left for later writers to do, except to abridge or excerpt, and this was quite in accord with the times in which they lived, when the sum total of historical, if not of all literary activity, lay in abridgments and excerpts. 1 The number rests on an emendation, though a sure one.